Genesi & PowerPC
  • ASiegel
    Posts: 1342 from 2003/2/15
    From: Central Europe
    Quote:

    Zylesea wrote:

    But even if it would stay below the PowerMac G5 performance in single core it would be smaller, new and energy efficient. Many (me included) don't buy into G5 kit because of fear of leaky cooling, energy consumption and the neat, but bulky case. A T2080 design could be attractive in this regard - given the price tag stays sane.

    I think it is worthwhile to go back and review feanorĀ“s research and conclusion. Pricing depends largely on how many units you believe could be sold, which is not going to be many...
  • »11.11.15 - 09:25
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 11631 from 2003/5/22
    From: Germany
    > Dunno how Ativec competes against VMX.

    If e6500's AltiVec implementation and integration are based on e600's (which I'm pretty sure they are), it should be faster per clock than PPC970's VMX.
  • »11.11.15 - 09:28
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    redrumloa
    Posts: 1423 from 2003/4/13
    Quote:

    Andreas_Wolf wrote:
    > I doubt single core performance of these phones. without extra hardware, would
    > dethrone the single core performance of a G5.

    https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=11328&forum=11&start=4

    > What is the FPU situation on a P8?

    Compliant with IEEE-754, so standard PPC FPU.

    > Does it have altivec?

    Yes, POWER series has it since POWER6. IBM calls it VMX, though.

    > IIRC A_W posted a link some time back to a Tyan based P8 system that was $4k.

    US$2.85k.

    https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=9463&start=76

    Someone said it was not really available for purchase, though:

    https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=9463&start=84


    All very good info. This is you at your best Andreas_Wolf, thanks!
  • »11.11.15 - 15:18
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    amigadave
    Posts: 2793 from 2006/3/21
    From: Northern Calif...
    Quote:

    Jim wrote:
    One other advantage to my last proposal?
    The e6500 cored T2080 features AltiVec instructions.


    I wonder what it would cost to get a T2080 motherboard designed and produced roughly, using as much community knowledge and manpower as possible to keep the cost down where ever possible?

    The T2080 seems to be the logical best choice for a low (but not lower than the SAM460cr) cost system which AmigaOS4.x and MorphOS could be ported to.

    I doubt we could create such a system, but I am curious what some of you think it would cost to actually complete.
    MorphOS - The best Next Gen Amiga choice.
  • »14.11.15 - 03:00
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    @amigadave

    A whole lot. And it won't make much sense now the SAM, X5000 and Tabor exist. The market is already full.
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »14.11.15 - 07:59
    Profile Visit Website
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4967 from 2009/1/28
    From: Delaware, USA
    Quote:

    Yasu wrote:
    @amigadave

    A whole lot. And it won't make much sense now the SAM, X5000 and Tabor exist. The market is already full.


    Well, from the DMIPS figures Andreas has provided a 1.8GHz T2080 would be almost twice as powerful as a 2.2GHz P5040 (which, in itself, is twice as powerful as the T5020).

    And that is with about 90% of the single threaded performance of the X5000.
    So, if a community funded board was approached, that is the CPU I'd like to see used.

    Its cheaper than the P50XX, and the performance (overall) would render it a superior choice to the systems you have mentioned.
    "Never attribute to malice what can more readily explained by incompetence"
  • »14.11.15 - 16:11
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 11631 from 2003/5/22
    From: Germany
    > from the DMIPS figures Andreas has provided a 1.8GHz T2080 would be almost
    > twice as powerful as a 2.2GHz P5040

    It's (1.8*8*3.0)/(2.2*4*3.0)-1=64% faster. But that's requiring SMP support anyway, which we don't have.

    > which, in itself, is twice as powerful as the T5020

    Assuming you mean the P5020, it's even (2.2*4)/(2.0*2)=2.2 times as fast.

    > that is with about 90% of the single threaded performance of the X5000.

    Yes, with X5000/40. Compared to X5000/20 it's (1.8*3.3)/(2.0*3.0)=99%.
  • »14.11.15 - 17:54
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    Yeah, but then we will have four boards competing for a couple of hundreds of buyers. That is not good prospects to get your development money back.
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »14.11.15 - 18:09
    Profile Visit Website
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4967 from 2009/1/28
    From: Delaware, USA
    Well, I did want to be conservative Andreas. ;)
    "Never attribute to malice what can more readily explained by incompetence"
  • »14.11.15 - 18:30
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    amigadave
    Posts: 2793 from 2006/3/21
    From: Northern Calif...
    If the T2080 provides 99% of the performance of the P5020 and 90% of the performance of the P5040, at a substantially lower price, PLUS has the e6500 core with Altivec, I hope that Trevor is reading this thread and drops the X5000 w/P5040 design, and instead switches to a lower priced, high end AmigaOne system using the T2080 CPU in the future.

    Of course, for that to happen, he probably needs to sell more than 500 Tabor motherboards or complete systems, to allow him to continue forward in producing PPC motherboards for the AmigaOS4.x and MorphOS users.

    With all the negative discussion regarding the Tabor's CPU choice, I am nervous about how well the Tabor will sell, and how much poor sales could hurt A-Eon's future. I just hope that once the Tabor is released and tested running AmigaOS4.1FE, it's performance will disprove most of the negative comments made about it, and the sale price will be low enough to spur many hundred sales, so that A-Eon can hopefully continue and make better choices in the future.

    Trevor, if you are reading this, please add one or two of the MorphOS Dev. Team members to your decision making team, before you commit large sums of money on your next motherboard design. Although I might criticize some of your decisions (using hind sight long after the fact), I really do want to see you succeed and continue to provide hardware and software for the whole Amiga community, long into the future.

    [ Edited by amigadave 15.11.2015 - 16:24 ]
    MorphOS - The best Next Gen Amiga choice.
  • »16.11.15 - 00:23
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    bbrv
    Posts: 750 from 2003/2/14
    From: Earth
    Quote:

    Andreas_Wolf wrote:
    That's obviously not what I'm talking about. I was referring to the chronology, i.e. that the involvement of the Genesi forerunner Thendic France started only when the Pegasos I design was completed. Genesi was involved when the Pegasos II was designed. This doesn't imply that Bill Buck had a hand in routing its board components, does it?

    DCE wasn't paid for the manufacturing but only for the lease of the mounting machines. It was bplan staff who operated them.


    ...the day before the Aachen "show" in 2002 (& Genesi Aachen gallery) Raquel and I ran the CPU cards through the wave solder machine after placing the capacitors on the boards. Gerald and Thomas ran the SMT line with Thomas Dellert. Thomas was paid for storing components at DCE, as well as for other services.

    As I recall, the Aachen event marked the transition from the Betatester to the Pegasos. There were a number of "shows" prior to this (scroll down to links in Products Overview & History -- Genesi Products). There are actually pictures here from that time period too at DCE, in Frankfurt and Paris, and in Normandy.

    We reconnected post-VisCorp with Gerald at the Amiga show in Cologne in 2000 (Amiga-News.de Interveiw with Petro). The Pegasos was called the "Betatester" before it shipped. There were a number of development challenges. The Mai Logic Articia being the most troublesome. It is fair to say we contributed significantly to the cost of the development and were involved before the Pegasos was a product.

    R&B
  • »16.11.15 - 01:41
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 11631 from 2003/5/22
    From: Germany
    > I hope that Trevor [...] drops the X5000 w/P5040 design, and instead switches
    > to a lower priced, high end AmigaOne system using the T2080 CPU in the future.

    The X5000/40 is planned to be released next year, so hardware development has most likely been completed by now. Judging by development cycle of prior A-Eon products, developing a new board based on T2080 would require another 2 years or more, so I'm not sure dropping the X5000/40 now would be beneficial to a future T2080-based A-Eon product.
  • »16.11.15 - 11:57
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 11631 from 2003/5/22
    From: Germany
    >> DCE wasn't paid for the manufacturing but only for the lease of the
    >> mounting machines. It was bplan staff who operated them.

    > Raquel and I ran the CPU cards through the wave solder machine after placing
    > the capacitors on the boards. Gerald and Thomas ran the SMT line with Thomas Dellert.
    > Thomas was paid for storing components at DCE, as well as for other services.

    Thanks for the additional info and for the confirmation that DCE wasn't commissioned to manufacture the boards.

    >> the involvement of the Genesi forerunner Thendic France started only
    >> when the Pegasos I design was completed.

    > the Aachen event marked the transition from the Betatester to the Pegasos. [...] The Pegasos
    > was called the "Betatester" before it shipped. There were a number of development challenges.
    > The Mai Logic Articia being the most troublesome. [...] we contributed significantly
    > to the cost of the development and were involved before the Pegasos was a product.

    Yes, I was obviously referring to the original post-prototype Pegasos board, i.e. pre-April. I know you were key in financing the development and production of the intermediate "April" PCB that was added to the Pegasos PCB before it became available for non-betatesters. However, I remember that also the pre-April board was called Pegasos, and "Betatester" was used more like a byname.

    http://www.amigahistory.plus.com/press/04102002.html

    By the way, if my command of the English language isn't completely awry, "Betatestee" would have made way more sense as byname of the board ;-)
  • »16.11.15 - 13:46
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    CountRaven
    Posts: 563 from 2007/12/11
    From: Greece
    We need fast and cost affordable PowerPC Genesi Hardware able to run MorphOS and if possible AmigaOS.
  • »17.11.15 - 01:00
    Profile Visit Website
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Intuition
    Posts: 1106 from 2013/5/24
    From: Nederland
    Perhaps crowd funding a new motherboard would be a good idea?

    I know I'd contribute to such a product depending on the specs and the official blessing of the MorphOS team obviously.
    1.67GHz 15" PowerBook G4, 1GB RAM, 128MB Radeon 9700M Pro, 64GB SSD, MorphOS 3.15

    2.7GHz DP G5, 4GB RAM, 512MB Radeon X1950 Pro, 500GB SSHD, MorphOS 3.9
  • »17.11.15 - 11:50
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    CountRaven
    Posts: 563 from 2007/12/11
    From: Greece
    Me too.
  • »17.11.15 - 13:39
    Profile Visit Website
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    That would be a problem as MorphOS Team doesn't support non existing boards. You need to make it first before they will start working on a port. Or so they keep saying at least.
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »17.11.15 - 16:22
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 11631 from 2003/5/22
    From: Germany
    > That would be a problem as MorphOS Team doesn't support non existing boards. You need
    > to make it first before they will start working on a port. Or so they keep saying at least.

    Would it even be feasible any other way?
  • »17.11.15 - 16:58
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4967 from 2009/1/28
    From: Delaware, USA
    Quote:

    CountRaven wrote:
    We need fast and cost affordable PowerPC Genesi Hardware able to run MorphOS and if possible AmigaOS.


    Did Bill say that the meeting scheduled with Freescale in December was as a representative of Genesi or power2people?

    Whether crowd or bounty funded, I would back an alternative board.

    Something that was powerful and priced reasonably.
    "Never attribute to malice what can more readily explained by incompetence"
  • »18.11.15 - 13:52
    Profile
  • ASiegel
    Posts: 1342 from 2003/2/15
    From: Central Europe
    Quote:

    Andreas_Wolf wrote:
    Yes, I was obviously referring to the original post-prototype Pegasos board, i.e. pre-April. I know you were key in financing the development and production of the intermediate "April" PCB that was added to the Pegasos PCB before it became available for non-betatesters. However, I remember that also the pre-April board was called Pegasos, and "Betatester" was used more like a byname.

    Actually, it referred to combinations of pre-release versions of the Pegasos as well as pre-release versions of MorphOS. Here is the original announcement text (added emphasis for clarity):

    "Announcing BETATESTER and Team Betatester!

    The time has come for Thendic-France/bplan to start selling machines. We are not going to call it the Pegasos for now. We are going to call the pre-release version of the Pegasos with the pre-release version of MorphOS for PPC: BETATESTER.

    MorphOS may not be ready for regular computer users, but it is ready for developers. You can purchase a BETATESTER for 1000 Euros (tax not included)."

    It is also worth nothing that the original BETATESTER priced at 1000 EUR was a 'complete system', not merely a mainboard with software.

    Quote:

    By the way, if my command of the English language isn't completely awry, "Betatestee" would have made way more sense as byname of the board ;-)

    See above. Initially, it was not used as a byname. Later on, it took on a different meaning and inspired the name of the MorphOS mailing lists, which were called "beta2" and "beta2-info" until 2004.
  • »18.11.15 - 18:55
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 11631 from 2003/5/22
    From: Germany
    > ["Betatester"] referred to combinations of pre-release versions of the Pegasos as well as
    > pre-release versions of MorphOS.

    Thanks for clarification.

    >> "Betatestee" would have made way more sense as byname of the board ;-)

    > Initially, it was not used as a byname.

    Yes, but byname or not, something that is to be tested is usually referred to as "testee", not "tester" :-)
  • »19.11.15 - 00:02
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    amigadave
    Posts: 2793 from 2006/3/21
    From: Northern Calif...
    Quote:

    Yasu wrote:
    That would be a problem as MorphOS Team doesn't support non existing boards. You need to make it first before they will start working on a port. Or so they keep saying at least.


    Since they have announced and actually begun work to move to x64 architecture, it makes no sense for them to become involved in more new PPC systems, but it is not beyond imagining that IF the MorphOS Dev. Team knew that it was going to take years to complete an architecture switch, and IF they also wanted to, they could help choose which components were to be used on a community funded, new motherboard design, and agree to support such a board when it was finished.

    That commitment could be stated prior to the board existing, but I doubt very much that such a plan will be put in place, and hopefully the switch to x64 will not take years to complete, counting from today, as work on it has already been begun some time in the past.

    Personally, I will be satisfied with a few releases which improve the performance and features of MorphOS on the hardware that is already supported, as we wait (hopefully only 6 to 18 months longer) for an x64 compatible version of MorphOS to be completed. If it is going to be significantly longer than that time frame, perhaps a newer, faster, and more modern, community funded PPC motherboard design would be worth while.

    Edit: I am also not supporting the idea of a community funded PPC motherboard at this point in time. I am hoping that the Tabor will be priced low enough, and perform well enough, so the MorphOS Dev. Team members will reconsider and reverse their decision NOT to support it. I doubt very much that there are enough AmigaOS4.x users to purchase 1000 Tabor systems and/or motherboards, but if MorphOS were ported to the Tabor, it would no doubt sell additional numbers of systems or motherboards. Once the Tabor is released for sale and we find out how well (or how poorly) it runs AmigaOS4.1FE, plus we know for sure how much it will cost, then I might have to re-evaluate my point of view regarding a community funded PPC motherboard.

    [ Edited by amigadave 18.11.2015 - 20:35 ]
    MorphOS - The best Next Gen Amiga choice.
  • »19.11.15 - 04:23
    Profile
  • ASiegel
    Posts: 1342 from 2003/2/15
    From: Central Europe
    Quote:

    Andreas_Wolf wrote:
    Yes, but byname or not, something that is to be tested is usually referred to as "testee", not "tester" :-)

    It depends on what exactly the main object of your testing is. The idea behind the BETATESTER was to get systems into the hands of eager early adopters so they could help to improve the included operating system by reporting issues. Put differently, the pre-release version of the Pegasos mainboard was primarily a tool for testing MorphOS. Much like a person uses a "cooker" as a tool to cook food, the "BETATESTER" hardware was used to (beta)test MorphOS.

    While the hardware was also given the "pre-release" label, I think it is reasonable to say that there was no general expectation that testers would have the required capabilities to correctly identify and report hardware bugs, which would have been extremely difficult since the hardware was running software that was known to be unfinished and be the most likely source of problems.
  • »19.11.15 - 08:20
    Profile
  • Order of the Butterfly
    Order of the Butterfly
    Everblue
    Posts: 166 from 2004/1/6
    Maybe this is a naive question... but instead of "crowdfunding a new motherboard" (for MorphOS in this case), isn't it easier to crowdfund MorphOS port for a particular x86 chosen by MorphOS Team?
  • »19.11.15 - 09:23
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    @Everblue

    I agree. We already have the SAM 460 and the upcoming X5000. We _really_ don't need any more PPC boards. They will probably never be cheap enough for me to get one of those instead of a used PPC Mac anyway.

    I much rather support the ISA shift. It will take it's time no matter what we do but anything to speed it up would be good.

    Another idea is to create a Patreon (donating a fixed sum every month) to a particular developer or to the entire team. I used to do this every month for AROS for almost 2 years. I think if we use Power2people it won't cost much, if any, fees.

    Yet another idea is to create a "beer-and-pizza" fund which the Team is free to use for that purpose (or coke and cookies if they prefer) when they code, individually or when they meet.

    It really depends on what they prefer, but the Team should not be intimidated by our "shut up and take our money" attitude. I know perfectly well I might not get my moneys worth back. And I don't care :-)
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »19.11.15 - 09:59
    Profile Visit Website